Assistant Professor Gabriele Farina mines the foundations of choice-making in complicated multi-agent situations.
Gabriele Farina grew up in a small town in a hilly winemaking place of northern Italy. Neither of his parents had college degrees, and although both had been convinced they “didn’t understand math,” Farina stated, they bought him the technical books he need and didn’t discourage him from attending the science-orientated, instead of the classical, high school.
By age 14, Farina had focused on an idea that might prove foundational to his career.
“I become interested very early via the idea that a machine may make predictions or choices a lot more better than humans,” he stated. “The truth that human-made mathematics and algorithms could form systems that, in some sense, surpass their creators, all at the same time as building on just building blocks, has usually been a primary source of awe for me.”
At age 16, Farina wrote code to solve a board game he played with his 13-year-old sister.
“I used game after game to compute the most advantageous move and show to my sister that she had already lost long earlier than both of us could see it ourselves,” Farina stated, including that his sister was less fascinated along with his new system.
Now an assistant professor in MIT’s Department of Electrical Engineering and Computer Science (EECS) and a principal investigator on the Laboratory for Information and Decision Systems (LIDS), Farina integrates ideas from game theory with such tools as machine learning, optimization, and statistics to advance theoretical and algorithmic foundations for decision-making.
Enrolling at Politecnico di Milano for college, Farina studied automation and control engineering. Over time, moreover, he noticed that what activated his interest was not “simply applying known techniques, but understanding and broadening their foundations,” he stated. “I progressively shifted more and more toward theory, even as still worrying deeply about showing concrete applications of that theory.”
Farina’s advisor at Politecnico di Milano, Nicola Gatti, professor and researcher in computer science and engineering, presented Farina to research inquiries in computational game theory and endorsed him to apply for a PhD. At the time, being the primary in his instant family to earn a college degree and living in Italy, in which doctoral degrees are treated in a different way, Farina stated that he didn’t even realize what a PhD was.
However, one month after graduating along with his undergraduate degree, Farina started a doctoral degree in computer science at Carnegie Mellon University. There, he won distinctions for his research and dissertation, as well as a Facebook Fellowship in Economics and Computation.
As he was completing his doctorate, Farina worked for a year as a research scientist in Meta’s Fundamental AI Research Labs. One of his major venture was supporting to develop Cicero, an AI that was able to beat human players in a game that entails forming alliances, negotiating, and detecting when other players are bluffing.
Farina stated that, “whilst we built Cicero, we designed it so that it would now not comply with alliance if it was not in its interest, and it likewise understood whether a player was possibly lying, because for them to do as they proposed would be against their own incentives.”
A 2022 article within the MIT Technology Review stated that Cicero should represent advancement towards AIs that can solve complex issues needing compromise.
After his year at Meta, Farina joined the MIT faculty. In 2025, he was distinguished with the National Science Foundation CAREER Award. His work — based on game theory and its mathematical language describing what takes place when distinct parties have exceptional goals, and then quantifying the “equilibrium” wherein no one has a reason to change their strategy— objectives to simplify huge, complicated real-world scenarios where calculating such an equilibrium could take a billion years.
“I research how we can use optimization and algorithms to actually locate those strong factors successfully,” he stated. “Our work tries to shed new light at the mathematical underpinnings of the theory, better manage and are expecting these complex dynamical systems, and uses these ideas to compute accurate solutions to huge multi-agent interactions.”
Farina is mainly interested in settings with “imperfect information,” this means that that some agents have information this is unknown to other participants. In such eventualities, information has value, and participants need to be strategic about acting on the information they own in order now not to expose it and reduce its value. An everyday example takes place in the game of poker, in which players bluff in order to conceal information about their cards.
As per Farina, “we now live in a world in which machines are far better at bluffing than humans.”
A situation with “large amounts of imperfect records,” has introduced Farina returned to his board-game starting’s. Stratego is a army strategy game that has inspired studies efforts costing millions of dollars to produce systems capable of beating human players. Needing complex risk calculation and misdirection, or bluffing, it was likelihood the only classical game for which major efforts had failed produce superhuman overall performance, Farina stated.
With new algorithms and training costing less than $10,000, instead of millions, Farina and his research team have been capable of beat the best players of all time — with 15 wins, four draws, and one loss. Farina stated that he is thrilled to have produced such outcomes so economically, and he hopes “those new strategies can be included into future pipelines,” he says.
“We have seen constant development closer to constructing algorithms that could purpose strategically and make sound choices despite huge action spaces or imperfect records. I am excited about seeing these algorithms integrated into the broader AI revolution that’s going on around us.”
